Sometimes Losing Leads to Winning

Seminal evidence from Berger and Pope (2011) shows that teams slightly behindat halftime are more likely to win. A recent study by Klein Teeselink et al. (2023),however, suggests this finding is an artifact confined to the 1999–2009 NBA seasons.Using an independently assembled dataset of 546,628 professional games, I revisitthis question with a local regression-discontinuity design. I document that the“behind-at-halftime advantage” is not an artifact: it appears robustly in NBA/ABAplay in the late 1960s–early 1970s and in non-U.S. leagues long after the originalstudy. Its attenuation in the modern NBA coincides with strategic adjustments byleading teams, who appear to have learned to counter their opponents’ motivationalsurge. The results show that reference-dependent motivation is a robust phenomenon,but its expression in equilibrium depends on context and strategic adaptation byopponents.